Article
I’m Sorry About the Rain!
Superfluous Apologies Demonstrate
Empathic Concern and Increase Trust
Social Psychological and
Personality Science
00(0) 1-8
ª The Author(s) 2013
Reprints and permission:
sagepub.com/journalsPermissions.nav
DOI: 10.1177/1948550613506122
spps.sagepub.com
Alison Wood Brooks1, Hengchen Dai2, and Maurice E. Schweitzer2
Abstract
Existing apology research has conceptualized apologies as a device to rebuild relationships following a transgression. Individuals,
however, often apologize for circumstances for which they are obviously not culpable (e.g., heavy traffic or bad weather). In this
article, we define superfluous apologies as expressions of regret for an undesirable circumstance for which the apologizer is clearly
not responsible. Across four studies, we find that superfluous apologies increase trust in the apologizer. This effect is mediated by
empathic concern. Issuing a superfluous apology demonstrates empathic concern for the victim and increases the victim’s trust in
the apologizer.
Keywords
superfluous apology, apology, trust, benevolence-based trust, empathy, stochastic trust game
Hi, folks. Well, I’m sorry about the rain.
President Bill Clinton (August 1995)
Apology research has presumed that the apologizer bears
responsibility for harming the victim. This presumption of
responsibility is embedded in existing conceptualizations and
definitions of apologies. According to Goffman (1971), an
apology splits an individual into two parts, ‘‘one half of the
individual representing the wrongdoing, and the other half . . .
hoping to be forgiven’’ (p.113); and apologies have been
defined as ‘‘admissions of blameworthiness and regret for an
undesirable event that allow actors to try to obtain a pardon
from audiences’’ (Goffman, 1971; Lazare, 2004; Ohbuchi,
Kameda, & Agarie, 1989; Scher & Darley, 1997; Schlenker,
1980; Schlenker & Darby, 1981, p. 271; Schweitzer, Hershey,
& Bradlow, 2006; Tavuchis, 1991). Incorporating this
definition, the extant literature has studied apologies as a post
hoc device for restoring relationships between a culpable apologizer and a target (Haselhuhn, Schweitzer, & Wood, 2010;
Kim, Dirks, Cooper, & Ferrin, 2006; Kim, Ferrin, Cooper, &
Dirk, 2004; Van Kleef, De Dreu, & Manstead, 2010).
We argue that this approach is overly narrow. Many individuals, such as Bill Clinton, apologize when they are clearly not
responsible. In contrast to the existing research that has assumed
that the apologizer is blameworthy, we investigate the use of
apologies for situations in which the apologizer is obviously not
at fault.
We redefine apologies as expressions of regret for an undesirable event or situation. This broader definition no longer
constrains apologies to include ‘‘admissions of blameworthiness’’
and does not presume that an apology aims to obtain ‘‘pardon
from audiences.’’ In particular, our broader definition includes
what we term superfluous apologies: expressions of regret for
an undesirable circumstance that is clearly outside of one’s
control.
Although prior work has assumed that apologies reflect
admissions of blameworthiness for causing harm, superfluous
apologies are common. For example, Japanese often apologize
even when they are not blameworthy (Maddux, Kim, Okumura,
& Brett, 2011). It is possible that people feel guilty when they
observe others suffering from an undesirable circumstance,
even when the observer is blameless. Prior work has demonstrated that culpability is not required to experience feelings
of guilt (e.g., Baumeister, Stillwell, & Heatherton, 1994).
Feelings of guilt in such contexts, separate from feelings of
responsibility, may be sufficient to motivate an apology.
By issuing a superfluous apology, the apologizer communicates that he has taken the victim’s perspective, acknowledges
adversity, and expresses regret. For example, though Bill
Clinton may have been dry and comfortable, by saying ‘‘I’m
sorry about the rain,’’ he communicates that he understands the
1
Harvard Business School, Negotiation, Organizations & Markets Unit, Boston,
MA, USA
2
Wharton School, University of Pennsylvania, PA, USA
Corresponding Author:
Alison Wood Brooks, Harvard Business School, Negotiation, Organizations &
Markets Unit, Baker Library 463, Boston, MA 02459, USA.
Email: awbrooks@hbs.edu
2
crowd’s perspective and expresses regret that they are wet and
uncomfortable in the rain. Previous work demonstrates that
negotiators who express regret or guilt are liked better than
those who do not (Van Kleef, De Dreu, & Manstead, 2006).
This finding is related to apology research that has found that
the act of apologizing represents an empathic expression that
‘‘demonstrates the offenders’ recognition of and concern for
their victims’ suffering’’ (Fehr & Gelfand, 2010, p. 38). We
fundamentally advance prior apology work by demonstrating
that superfluous apologies demonstrate empathy even in the
absence of culpability.
In our investigation, we expect superfluous apologies to
demonstrate empathic concern and consequently increase
benevolence-based trust, the extent to which we believe
someone is kind and concerned about our well-being (Johnson,
Cullen, Sakano, & Takenouchi, 1996; Mayer & Davis, 1995).
Benevolence-based trust correlates strongly with liking and
behavioral trust.
Although we anticipate several perceptual and behavioral
benefits of superfluous apologies, some prior work suggests
that apologizing may have drawbacks (Bergsieker, Shelton,
& Richeson, 2010). For example, Tannen (1996, 1999, 2001)
conjectures that apologies may harm perceived power and
competence, especially for women. Consequently, in addition
to measuring empathy and benevolence-based trust, we also
explore how superfluous apologies influence perceptions of
power and competence-based trust, the perception that someone possesses the ability and skills required for a task (Butler
& Cantrell, 1984; Kim et al., 2004).
Study 1
In Study 1, we investigate the effect of superfluous apologies
on trust. In this study, we measure behavioral trust using a
novel version of the trust game, the stochastic trust game, and
we collect attitudinal measures of trust and liking.
Method
Participants
We recruited 178 students from a Northeastern university to
participate in an experiment for pay (60.0% female, Mage ¼
19.98, standard deviation [SD] ¼ 2.46). Participants received
a US$10 show up fee and had the opportunity to earn additional
compensation of up to US$9 based on the performance.
Design and Procedure
We randomly assigned participants to one of two conditions
(superfluous apology vs. no apology). Our primary dependent
variable was behavior in a repeated trust game.
We recruited participants to a behavioral laboratory. We
informed participants that they would be playing several
rounds of a repeated trust game with an anonymous counterpart
in the room (e.g., Berg, Dickhaut, & McCabe, 1995; Murphy,
Social Psychological and Personality Science 00(0)
Rapoport, & Parco, 2000). In reality, all participants played the
same role against a common, computer-simulated counterpart.
We informed participants that they would be endowed with
US$6 in each round, which they could either pass to their
counterpart or keep. If they chose to pass the US$6 to their
counterpart, the money would be tripled (to US$18). The
counterpart could then either keep the US$18 or pass half of the
money (US$9) back. Consistent with prior work, we operationalized trust as participants’ decision to pass their endowment
(Bohnet & Huck, 2004; Engle-Warnick & Slonim, 2004;
Haselhuhn et al., 2010; Lount & Pettit, 2012; Malhotra & Murnighan, 2002).
We explained to participants that both players would make
decisions simultaneously and that players would learn about
their counterpart’s decision regardless of what their counterpart
chose. For example, if participants chose to keep their endowment, they would still learn whether or not their counterpart
would have returned US$9.
Importantly, we explained that there was a 25% chance that
the computer would override the counterpart’s decision on each
round, but the computer would never override the counterpart’s
decision on the last round. We term this version of the game the
stochastic trust game. We informed participants that, after each
round, we would reveal both players’ decisions, whether the
computer overrode their partner’s decision, and the outcome
of the round.
All participants completed four rounds of the trust game. Both
the decision of their counterpart and the purportedly random overriding action of the computer were preprogrammed. In Rounds 1
and 2, participants learned that their counterpart chose to return
US$9 and that the computer did not override this decision. In
Round 3, participants learned that their counterpart chose to
return US$9 and that the computer overrode this decision.
After receiving feedback from Round 3 and before making
their decision for Round 4, half of the participants received the following superfluous apology message from their counterpart: ‘‘I’m
really sorry the computer changed my choice last round.’’ In the
no apology condition, participants did not receive a message.
Before Round 4, we announced that everyone would know
that this was the last round. The decision to pass in the final
round was our primary dependent variable. The decision to
pass in the final round, after the endgame has been announced,
represents the best measure of trust (cf. Bohnet & Huck, 2004;
Engle-Warnick & Slonim, 2004; Haselhuhn et al., 2010;
Malhotra & Murnighan, 2002; Schweitzer et al., 2006).
After the trust game, participants rated their partner with
respect to competence-based trust (3 items, a ¼ .91),
benevolence-based trust (3 items, a ¼ .83), likeability (10
items, a ¼ .88), and perceived power (for measure details refer
online supplemental material found at http://spps.sagepub.com/
supplemental).
Results
We found no main effects or interaction effects for gender, and
we report all of our findings collapsed across the gender.
Brooks et al.
3
85
77%
Passing Rate (%)
80
75
65%
70
65
60
55
Superfluous Apology
No Apology
Experimental Condition
Figure 1. Superfluous apology increases trusting behavior in a
repeated trust game (Study 1).
in the superfluous apology condition reported significantly
higher benevolence-based trust in their counterpart (M ¼
3.54, SD ¼ 0.68) than did participants in the no apology
condition (M ¼ 3.02, SD ¼ 0.85), t(161) ¼ 4.35, p < .001.
We did not find a significant effect of experimental condition
on perceptions of competence-based trust (p ¼ .23).
Perceptions of benevolence-based trust mediated the relationship between superfluous apology and trusting behavior
(Baron & Kenny, 1986). By including benevolence-based trust
in our model, the influence of receiving a superfluous apology
was no longer significant (b ¼.09, p ¼ .35) but perceptions of
benevolence-based trust predicted trusting behavior (b ¼ .40,
p < .001). Because our dependent variable was binary, we ran
a 5000-sample bootstrap (Preacher & Hayes, 2004, 2008) using
MacKinnon and Dwyer’s (1993) logistic regression method
and found a standardized indirect effect of.17 (SE ¼ 0.05,
95% biased-correct CI ¼ [0.08, 0.23]). We depict this mediation effect in Figure 2.
Perceptions of
benevolence-based trust
.40***
.34***
Trusting behavior
Superfluous apology
.09 (.24*)
* p < .05, ** p < .01, *** p < .001
Figure 2. Perceptions of benevolence-based trust mediate the relationship between superfluous apology and trusting behavior (Study 1).
Values are standardized regression coefficients. The value in parentheses represents the total effect of superfluous apology condition
on trusting behavior.
Trusting Behavior
Participants were more likely to pass US$6 to their counterpart
in the last round when they received a superfluous apology than
when they did not receive a superfluous apology: 77% (65 of 84)
of participants in the apology condition passed US$6 compared
to 65% (51 of 79) of participants in the no apology condition. We
conducted a logistic regression with experimental condition as
the independent variable, passing behavior in the first three
rounds as a covariate to account for baseline trust,1 and final
round passing behavior as the dependent variable. We found a
significant effect of superfluous apology on trusting behavior,
b ¼ .24, B ¼ 1.11, standard error (SE) ¼ .45, w2(1, N ¼ 163)
¼ 6.10, p ¼ .01. We depict this result in Figure 1. We also found
that trusting behavior in the first three rounds significantly predicted passing rate in the last round, b ¼ .59, B ¼ 3.88, SE ¼ .64,
w2(1, N ¼ 163) ¼ 37.33, p < .001.
Perceived Trust
Consistent with our prediction, receiving a superfluous apology
increased perceptions of benevolence-based trust. Participants
Likeability and Power
Participants in the superfluous apology condition liked their
counterpart significantly more (M ¼ 3.60, SD ¼ 0.47) than did
participants in the no apology condition (M ¼ 3.28, SD ¼ 0.54),
t(161) ¼ 4.15, p < .001. We did not find a significant effect of
receiving a superfluous apology on the perceived power of the
counterpart (p ¼ .89).
Discussion
In Study 1, participants who received a superfluous apology
were more trusting than were participants who did not receive
an apology. This is true for both our attitudinal measure of
benevolence-based trust and our behavioral measure of passing
decisions in the trust game. Perceptions of benevolence-based
trust fully mediated the effects of superfluous apologies on
trusting behavior and liking toward the counterpart.
Study 2
In Study 2, we extend our investigation by considering
alternative explanations. In Study 1, it is possible that our control condition decreased benevolence-based trust. In the control
condition in Study 1, the confederate sent no message. This
might have seemed impolite. To investigate this alternative
explanation, we include two different comparison conditions
in Study 2. We include a traditional apology (‘‘I’m sorry to
interrupt.’’) and neutral greeting (‘‘How are you?’’).
In addition, we extend our investigation by exploring
superfluous apologies in a new domain. In Study 1, both the
unfortunate circumstance (the random action of the computer)
and the outcome (behavioral measure of trust) were related to
the trust game. In Study 2, the apology (for a flight delay) is
unrelated to our measure of trust (lending a cell phone). We
expect that an apology for a flight delay will increase trust even
when the subsequent interaction is unrelated to the flight.
4
Design and Procedure
We randomly assigned participants to one of the three betweensubject conditions (superfluous apology vs. traditional apology
vs. neutral greeting). In all three conditions, we asked participants to imagine that they were waiting for a flight to visit a
friend whose father recently passed away when the airline
announces that the flight has been delayed for 2 hours. We asked
participants to imagine that after they sat down in an adjacent
boarding area, a passenger from a different, nondelayed flight
approached them (full scenario included in online supplemental
material found at http://spps.sagepub.com/supplemental).
Participants then watched a short video in which a confederate, playing the role of the other passenger, approaches and
asks to borrow their cell phone. The confederate makes this
request in one of the following three ways: ‘‘Hi. I am sorry your
flight is delayed. Can I borrow your cell phone?’’ (superfluous
apology), ‘‘Hi. I am sorry to interrupt. Can I borrow your cell
phone?’’ (traditional apology), or ‘‘Hi. How are you? Can I borrow your cell phone?’’ (neutral greeting). Participants then
rated the confederate on measures of benevolence-based (a ¼
.87) and competence-based trust (a ¼ .70).
We use the decision to lend a stranger a cell phone as our
measure of trust. Cell phones are moderately expensive, small,
fragile, and often contain private information. Handing a cell
phone to a stranger makes the individual vulnerable and represents an act of trust.
Results
We found no main effects or interaction effects of gender, and
we report all of our findings collapsed across gender.
Perceived Trust
We conducted a one-way analysis of variance (ANOVA) with
perceptions of benevolence-based trust as the dependent
variable and experimental condition (superfluous apology vs.
traditional apology vs. neutral) as the independent variable.
Supporting our thesis, we found differences across the three
conditions, F(2, 174) ¼ 11.81, p < .0001, Z2p ¼ .12. Planned
contrasts revealed that the confederate was rated higher in
benevolence-based trust following the superfluous apology
(M ¼ 4.28, SD ¼ 1.15) than following the neutral greeting
(M ¼ 3.34, SD ¼ 1.13, t(174) ¼ 4.85, p < .0001). The confederate was also rated higher in benevolence-based trust following the superfluous apology than following the traditional
apology (M ¼ 3.88, SD ¼ 0.87, t(174) ¼ 2.06, p < .05). Further, consistent with past research showing that apologies can
4.8
4.28
Benevolence-based Trust
We recruited 177 U.S. residents (Mage ¼ 27.69 years, SD ¼
9.44; 30% female) to complete a 3-min online survey in
exchange for US$.30.
A
4.4
3.88
4
3.34
3.6
3.2
2.8
Superfluous Apology Conventional Apology
Neutral
Experimental Condition
B
Benevolence-based Trust
Method
Participants
Social Psychological and Personality Science 00(0)
5
4.59
4.6
3.91
4.2
4.03
3.8
3.4
3
Superfluous Apology Acknowledgment
Neutral
Experimental Condition
Figure 3. Superfluous apology increases perceptions of benevolencebased trust (Studies 2 and 3).
serve as an interpersonal lubricant (Lazare, 2004; Tavuchis,
1991), we found that perceptions of benevolence-based trust
were significantly higher in the traditional apology condition
than in the neutral greeting condition, t(174) ¼ 2.75, p < .01.
We depict this pattern of results in Figure 3 (Panel A).
We also found a significant effect of the experimental
condition on perceptions of competence-based trust, F(2,
174) ¼ 4.16, p ¼ .02, Z2p ¼ .05. However, as in Study 1, we did
not find evidence that issuing a superfluous apology diminished perceptions of competence-based trust. Instead, we found
that, compared to the neutral greeting condition (M ¼ 4.63,
SD ¼ 0.97), perceptions of competence-based trust were
significantly higher in both the superfluous apology condition,
M ¼ 5.05, SD ¼ 0.92, t(174) ¼ 2.42, p ¼ .02, and the traditional
apology condition, M ¼ 5.09, SD ¼ 0.97, t(174) ¼ 2.59,
p ¼ .01. The latter two conditions did not significantly differ
from each other (p ¼ .22).
Discussion
Participants rated a confederate to be higher in benevolencebased trust when the confederate used a superfluous apology
(‘‘I’m sorry your flight was delayed.’’) than when the confederate used a polite greeting (‘‘How are you?’’) or a traditional
apology (‘‘I’m sorry to interrupt.’’). This finding extends our
investigation to a different domain and a different operationalization of trust.
Brooks et al.
5
Method
Participants
contrasts revealed that the seller who apologized for the rain
engendered higher benevolence-based trust (M ¼ 4.59, SD ¼
1.00) than did both the seller in the acknowledgment condition,
M ¼ 3.91, SD ¼ 1.00, t(307) ¼ 5.13, p < .001, and the seller in
the neutral condition, M ¼ 4.03, SD ¼ 0.86, t(307) ¼ 4.23, p <
.001. We found no difference in benevolence-based trust
between the acknowledgment and neutral conditions (p ¼ .39).
We depict this pattern of results in Figure 3 (Panel B).
As in Study 1, we did not find a significant effect of experimental condition on perceptions of competence-based trust,
F(2, 307) ¼ .36, p ¼ .70, Z2p ¼ .002, and none of the three pairwise comparisons were significant.
We recruited 310 U.S. residents (Mage ¼ 32.33 years, SD ¼
12.14; 52% female) to complete an online survey in exchange
for US$.40.
Likeability
Study 3
In Study 3, we investigate the psychological mechanism that
links superfluous apologies with trust. We expect superfluous
apologies to reflect empathic concern. By saying ‘‘I’m sorry,’’
the superfluous apologizer acknowledges an unfortunate
circumstance, takes the victim’s perspective, and expresses
empathy for the negative circumstance, even though it is outside of his or her control.
Design and Procedure
We randomly assigned participants to one of the three betweensubject conditions (superfluous apology vs. acknowledgment
vs. neutral). In all three conditions, we asked participants to
imagine that they were about to purchase a used iPod on
Craigslist. On their way to meet the seller, we asked participants to imagine that they got caught in the rain (full scenario
included online supplemental material found at http://spps.
sagepub.com/supplemental). When they meet the seller, the
seller greets them in one of the following three ways: ‘‘Hi there.
Oh, I’m so sorry it’s raining.’’ (superfluous apology), ‘‘Hi there.
Oh, it’s raining.’’ (acknowledgment), or ‘‘Hi there.’’ (neutral).
After reading the scenario, participants rated the seller on
measures of benevolence-based trust (a ¼ .86) and
competence-based trust (a ¼ .72) as well as likeability (‘‘The
seller is likeable’’). We also adapted 7 items from Davis
(1983) to assess perceptions of empathic concern (a ¼ .88; see
online supplemental material found at http://spps.sagepub.com/
supplemental for details). A factor analysis with a varimax
rotation validated that our measures of benevolence-based
trust, competence-based trust, and empathic concern loaded
onto three factors (using eigenvalue ¼ 1). The three composite
scores indicate that our measures effectively captured the
intended constructs.
Results
We found no main effects or interaction effects for gender, and
we report all of our findings collapsed across gender.
Perceived Trust
The superfluous apology increased perceptions of benevolencebased trust. We conducted a one-way ANOVA with perceptions
of benevolence-based trust as the dependent variable and experimental condition (superfluous apology vs. acknowledgment vs.
neutral) as the independent variable. Consistent with our prediction, there were significant differences among the three conditions, F(2, 307) ¼ 15.06, p < .001, Z2p ¼ .09. Planned
The experimental conditions significantly influenced perceptions of likeability, F(2, 307) ¼ 8.61, p < .001, Z2p ¼ .05.
Planned contrasts revealed that the pattern of likeability
mirrored that of benevolence-based trust: Participants in the
superfluous apology condition liked the seller significantly
more (M ¼ 4.39, SD ¼ 1.23) than did participants in both the
acknowledgment condition, M ¼ 3.76, SD ¼ 1.16, t(307) ¼
3.97, p < .001, and the neutral condition, M ¼ 3.91, SD ¼
1.04, t(307) ¼ 3.00, p ¼ .003. The extent to which the seller
was perceived to be likeable did not differ in the latter two
conditions (p ¼ .35).
Empathic Concern
Perceptions of empathic concern varied significantly across
conditions, F(2, 307) ¼ 20.69, p < .001, Z2p ¼ .12. Consistent
with our prediction, planned contrasts revealed that perceived
empathic concern was higher in the superfluous apology condition (M ¼ 4.27, SD ¼ 0.88) than in both the acknowledgment
condition, M ¼ 3.57, SD ¼ 0.94, t(307) ¼ 5.93, p < .001, and
the neutral condition, M ¼ 3.66, SD ¼ 0.73, t(307) ¼ 5.10, p <
.001. Perceptions of empathic concern did not differ in the latter two conditions (p ¼ .44).
Mediation
Empathic concern mediated the effect of superfluous apology
on perceptions of benevolence-based trust. The acknowledgment condition and the neutral condition were not significantly
different on any measures, so we collapsed across these two
conditions in our mediation analyses. When we included
empathy in our model, the effect of superfluous apology was
no longer significant (from b ¼ .30, p < .001 to b ¼ .06, p ¼
.19) and the effect of empathy remained significant (b ¼ .70,
p < .001). A 5000-sample bootstrap test estimated a standardized indirect effect of .24 (SE ¼ .038, 95% biased-corrected
CI [0.16, 0.31]), indicating full mediation.2 We depict this
mediation effect in Figure 4.
6
Social Psychological and Personality Science 00(0)
Perceptions of
empathic concern
.34***
.70***
Superfluous apology
.06 (.30***)
Perceptions of
benevolence-based trust
* p < .05, ** p < .01, *** p < .001
Figure 4. Empathic concern mediates the relationship between
superfluous apology and increased perceptions of benevolence-based
trust (Study 3). Values are standardized regression coefficients. The
value in parentheses represents the total effect of superfluous apology
condition on perceptions of benevolence-based trust.
Discussion
In Study 3, we contrasted a superfluous apology with an
acknowledgment of the adverse circumstance (in this case,
rain). We found that individuals rated the seller who issued a
superfluous apology higher in benevolence-based trust and
likeability than the seller who acknowledged the bad weather
but did not apologize for it and the seller who made no mention
of the bad weather. Supporting our thesis, empathic concern
fully mediated the effect of the superfluous apology on perceptions of benevolence-based trust.
Study 4
In Studies 1–3, we found a consistent boost in trust following a
superfluous apology in both lab- and Web-based settings. In
Study 4, we extend our investigation to a field setting. We investigated the consequences of issuing a face-to-face superfluous
apology followed by a request to borrow a stranger’s cell phone.
Method
A male confederate who was blind to our hypotheses
approached 65 strangers (30 female, 35 male), one at a time,
in a large Northeastern train station and requested to borrow
their cell phone. We depict the experimental setup in Figure 5.
We conducted this study over the course of 2 days in November
2010 when the forecast called for steady rain. The accumulated
precipitation during each of the 1-hour sessions was equivalent
(0.06 inches of rain)3; the American Meteorological Society characterizes this level of precipitation as ‘‘moderate rain.’’
The confederate alternated the script he used across the 65
individuals he approached. Half of the time the confederate delivered a superfluous apology script, ‘‘I’m so sorry about the rain!
Can I borrow your cell phone?’’ Half of the time the confederate
delivered a neutral script, ‘‘Can I borrow your cell phone?’’
We recorded the date, time, gender of the participant, and
whether or not the participant handed their cell phone to the
confederate. We used the decision to hand over a cell phone
to the confederate as our behavioral measure of trust.
Figure 5. The confederate requests to borrow a stranger’s cell phone
in the train station (Study 4).
Results and Discussion
Of the 65 participants, 18 (28%) volunteered to give their cell
phone to the confederate. We conducted a logistic regression
with trusting behavior as the dependent variable and apology
condition, gender, and day as independent variables. Supporting our thesis, 47% (15 of 32) of the participants in the
superfluous apology condition entrusted their cell phone to the
confederate, compared to 9% (3 of 33) of the participants in the
neutral condition, b ¼ .66, B ¼ 2.83, SE ¼ 1.15, w2(1, N ¼ 65)
¼ 5.99, p ¼ .01. We did not find main effects of gender (p ¼
.65), day (p ¼ .52), or interaction effects between apology
condition and day (p ¼ .46) on trusting behavior. We depict
these results in Figure 6.
Results from this field study reveal that a face-to-face superfluous apology increases trusting behavior. Strangers were
more likely to hand their cell phone to a confederate when the
confederate apologized for the rain than when he did not.
General Discussion
Extant research has conceptualized apologies as expressions of
remorse by culpable individuals to repair relationships and
restore lost trust. We challenge this paradigm. We broaden the
conceptualization of apologies to include expressions of
Brooks et al.
Cell Phone Lending Rate (%)
60
7
47%
50
40
30
20
9%
10
0
Superfluous Apology
Some scholars have suggested that there may be drawbacks
to issuing an apology. We did not identify any drawbacks in our
investigation. Still, it is quite possible that the repeated use of
superfluous apologies or the delivery of a superfluous apology
that appears insincere may yield different results.
Across our studies, we identify significant benefits to
apologizing. Superfluous apologies represent a powerful and
easy-to-use tool for social influence. Even in the absence of
culpability, individuals can increase trust and liking by saying
‘‘I’m sorry’’—even if they are merely ‘‘sorry’’ about the rain.
No Apology
Experimental Condition
Figure 6. Superfluous apology increases trusting behavior in a train
station (Study 4).
remorse by individuals who are obviously not culpable for a
transgression (e.g., ‘‘I’m sorry about your loss, the rain, bad
traffic’’). We call this type of apology a superfluous apology.
Using multiple methodologies across four studies, we document that superfluous apologies increase benevolence-based
and behavioral trust. The relationship between superfluous
apologies and trust is mediated by perceptions of empathic concern. In a laboratory study with a behavioral measure of trust,
individuals were more likely to pass money to confederate
counterparts who issued a superfluous apology than to those
who did not. In a video-vignette study, a superfluous apology
increased benevolence-based trust relative to both a commonly
used greeting and a traditional apology. In a vignette study, we
find that perceptions of empathic concern mediate the effect of
superfluous apologies on trust. Finally, in a field study, we find
that individuals were more likely to hand over their cell phone
to a stranger in a train station (our measure of trust) when they
received a superfluous apology than when they did not.
Although superfluous apologies increased trust in all of our
studies, a number of factors may moderate the relationship
between superfluous apologies and trust. For example, characteristics of the apologizer’s delivery (e.g., frequency, sincerity),
internal states of the apology recipient, such as emotion (Dunn
& Schweitzer, 2005), or characteristics of the apologizer or
apology recipient, such as status (Lount & Pettit, 2012), may
limit the effectiveness of superfluous apologies. For example,
Van Kleef and De Dreu (2010) found that prosocial individuals
responded generously but selfish negotiators responded exploitatively when they received an apology during a negotiation.
In our studies, participants knew that the apologizer was
clearly not responsible for the unfortunate circumstances
participants’ faced. In practice, however, there is a continuum
of culpability. This is true with respect to both the magnitude of
culpability and the certainty of blameworthiness. Recent work
by Gunia (2013) suggests that in ambiguous situations, when it
is unclear who is to be blamed, people reward those who take
blame more than those who express remorse, are evasive, or
deny responsibility. It is possible that the effects we observe
in these studies would be even stronger when culpability is
ambiguous.
Acknowledgments
The authors are grateful for feedback from John C. Hershey, Adam
Galinsky, Katherine Milkman, and Brian Gunia as well as valuable
research assistance from Tim Flank and Jessica Chen.
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to
the research, authorship, and/or publication of this article.
Funding
The authors thank the Wharton Ackoff Fellowship and the Wharton
Behavioral Lab for research support.
Notes
1. We included passing decision in earlier rounds as a covariate to
account for baseline trust. Without this control, the difference
between the two experimental conditions was marginally significant (p ¼ .07).
2. We also conducted a mediation analysis without collapsing the
neutral and acknowledgment conditions (Preacher & Hayes,
2004). When we compared the superfluous apology condition with
the acknowledgement condition, a 5000-sample bootstrap showed
a standardized indirect effect of 0.27 (standard error ¼ .05, 95%
bias-corrected CI [0.18, 0.36]). When we compared the superfluous
apology condition with the neutral condition, a 5000-sample bootstrap showed a standardized indirect effect of 0.22 (standard error
¼ 0.04, 95% bias-correct CI [0.15, 0.32]).
3. Weather data retrieved from http://cdo.ncdc.noaa.gov/qclcd/
QCLCD
Supplemental Material
The online [appendices/data supplements/etc] are available at http://
XXX.sagepub.com/supplemental.
References
Baron, R. M., & Kenny, D. A. (1986). The moderator-mediator variable distinction in social psychological research: Conceptual, strategic and statistical considerations. Journal of Personality and
Social Psychology, 51, 1173–1182.
Baumeister, Stillwell, & Heatherton. (1994). Guilt: An interpersonal
approach. Psychological Bulletin, 115, 243–267.
Berg, J., Dickhaut, J., & McCabe, K. (1995). Trust, reciprocity, and
social history. Games and Economic Behavior, 10, 122–142.
Bergsieker, H. B., Shelton, J. N., & Richeson, J. A. (2010). To be liked or
respected: Whites’ and minorities’ divergent goals during interracial
8
interactions. Journal of Personality and Social Psychology, 99,
248–264.
Bohnet, I., & Hucks, S. (2004). Repetition and reputation: Implication
for trust and trustworthiness when institutions change. American
Economics Review, 94, 362–366.
Butler, J. K., & Cantrell, R. S. (1984). A behavioral decision theory
approach to modeling dyadic trust in superiors and subordinates.
Psychological Reports, 55, 19–28.
Davis, M. H. (1983). Measuring individual differences in empathy:
Evidence for a multidimensional approach. Journal of Personality
and Social Psychology, 44, 113–126.
Dunn, J., & Schweitzer, M. (2005). Feeling and believing: The
influence of emotion on trust. Journal of Personality and Social
Psychology, 88, 736–748.
Engle-Warnick, J., & Slonim, R. L. (2004). The evolution of strategies
in a repeated trust game. Journal of Economics Behavior and
Organization, 55, 553–573.
Fehr, R., & Gelfand, M. (2010). When apologies work: How matching
apology components to victims’ self-construals facilitates forgiveness.
Organizational Behavior and Human Decision Processes, 113, 37–50.
Goffman, E. (1971). Relations in public: Microstudies of the public
order. New York, NY: Basic Books.
Grandey, A. A. (2003). When ‘‘the show must go on’’: Surface and deep
acting as determinants of emotional exhaustion and peer-rated service delivery. Academy of Management Journal, 46, 86–96.
Gunia, B. (2013). The blame bias: Forgoing the benefits of blametaking. Organization Science Working paper.
Haselhuhn, M. P., Schweitzer, M. E., & Wood, A. M. (2010). How
implicit beliefs influence trust recovery. Psychological Science,
21, 645–648.
Johnson, J. L., Cullen, J. B., Sakano, T., & Takenouchi, H. (1996). Setting the stage for trust and strategic integration in Japanese-U.S.
cooperative alliances. Journal of International Business Studeis,
27, 981–1004.
Kim, P. H., Dirks, K. T., Cooper, C. D., & Ferrin, D. L. (2006). When
more blame is better than less: The implications of internal vs.
external attributions for the repair of trust after a competencevs. integrity-based trust violation. Organizational Behavior and
Human Decision Processes, 99, 49–65.
Kim, P. H., Ferrin, D. L., Cooper, C. D., & Dirks, K. T. (2004).
Removing the shadow of suspicion: The effects of apology versus
denial for repairing competence- versus integrity-based trust violations. Journal of Applied Psychology, 89, 104–118.
Lazare, A. (2004). On apology. New York, NY: Oxford University Press.
Lount, R. B., Jr., & Pettit, N. C. (2012). The social context of trust:
The role of status. Organizational Behavior and Human Decision
Processes, 117, 15–23.
MacKinnon, D. P., & Dwyer, J. H. (1993). Estimating mediated
effects in prevention studies. Evaluation Review, 17, 144–158.
Maddux, W. W., Kim, P. H., Okumura, T., & Brett, J. M. (2011).
Cultural differences in the function and meaning of apologies.
International Negotiation, 16, 405–425.
Social Psychological and Personality Science 00(0)
Malhotra, D., & Murnighan, J. K. (2002). The effects of contract on
interpersonal trust. Administrative Science Quarterly, 47, 534–559.
Mayer, R. C., & Davis, F. D. (1995). An integration model of organization trust. Academy of Management Review, 20, 709–734.
Murphy, R. O., Rappaport, A., & Parco, J. E. (2006). The breakdown
of cooperation in iterative real-time trust dilemmas. Experimental
Economics, 9, 147–166.
National Archives and Records Administration, Bill Clinton’s Yellowstone National Park, August 1995 (National Archives and
Records Administration).
Ohbuchi, K., Kameda, M., & Agarie, N. (1989). Apology as aggression control: Its role in mediating appraisal of and response to
harm. Journal of Personality and Social Psychology, 56, 219–227.
Preacher, K. J., & Hayes, A. F. (2004). SPSS and SAS procedures
for estimating indirect effects in simple mediation models.
Behavior Research Methods, Instruments & Computers, 36,
717–731.
Preacher, K. J., & Hayes, A. F. (2008). Asymptotic and resampling
strategies for assessing and comparing indirect effects in multiple
mediator models. Behavior Research Methods, 40, 879–891.
Scher, S. J., & Darley, J. M. (1997). How effective are the things people say to apologize? Effects of the realization of the apology
speech act. Journal of Psycholinguistic Research, 26, 127–140.
Schlenker, B. R. (1980). Impression management: The self-concept,
social identity, and interpersonal relations. Monterey, CA:
Brooks/Cole.
Schlenker, B. R., & Darby, B. W. (1981). The use of apologies in
social predicaments. Social Psychology Quarterly, 44, 271–278.
Schweitzer, M. E., Hershey, J. C., & Bradlow, E. T. (2006). Promises
and lies: Restoring violated trust. Organizational Behavior and
Human Decision Processes, 101, 1–19.
Tannen, D. (1996). I’m sorry, I won’t apologize. Retrieved December
18, 2010, from https://www9.georgetown.edu/faculty/tannend/
nyt072196.htm
Tannen, D. (1999, April/May). Contrite makes right. Civilization, 6,
67–70.
Tannen, D. (2001). You just don’t understand: Women and men in
conversation. New York, NY: William Morrow.
Tavuchis, N. (1991). Mea culpa: A sociology of apology and reconciliation. Stanford, CA: Stanford University Press.
Van Kleef, G. A., & De Dreu, C. K. W. (2010). Longer-term consequences of anger expression in negotiation: Retaliation or
spillover? Journal of Experimental Social Psychology, 5,
753–760.
Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2006).
Supplication and appeasement in conflict and negotiation: The
interpersonal effects of disappointment, worry, guilt, and regret.
Journal of Personality and Social Psychology, 91, 124–142.
Van Kleef, G. A., De Dreu, C. K. W., & Manstead, A. S. R. (2010). An
interpersonal approach to emotion in social decision making: The
emotions as social information model. Advances in Experimental
social Psychology, 42, 45–96.